A few weeks ago, someone on Reddit did the thing nobody actually does they read LinkedIn’s algorithm guidelines. Not a summary. Not a thread of “hot takes” from a growth bro with 47 followers and a Canva carousel addiction. The actual guidelines. The engineering blogs. The community policies.

What they found should make you uncomfortable.

Because everything LinkedIn is now rewarding? It’s the opposite of what most founders are doing.

And everything LinkedIn is now punishing? It’s probably in your current strategy.

LinkedIn’s algorithm isn’t just scoring your content anymore. It’s scoring YOU.

The platform has rolled out AI models that evaluate creators not just individual posts. Your history matters. Your engagement patterns matter. Whether your audience is real or manufactured? That matters most of all.

Which means the game has changed. And if you’re still playing the 2023 version, you’re losing.

Engagement pods are dead. Not dying. Dead.

LinkedIn now detects cross-industry engagement patterns.

If a fitness coach, a crypto trader, and a SaaS founder are all liking each other’s posts within 90 seconds of publishing LinkedIn knows. Flags it as inauthentic. Throttles the reach. In some cases, permanently.

The logic is simple.

Your actual audience doesn’t behave like that. Real humans don’t sprint to your post the moment it drops, leave a generic “Great insight!” comment, and vanish.

Pods do. And LinkedIn’s systems have caught on.

A YC-backed founder someone with real credentials, real traction, real things to say had their account permanently banned after using Taplio. Not suspended. Banned. Gone.

A tool they probably bought to “save time” and “grow faster” deleted their entire professional presence overnight.

That’s not a cautionary tale. That’s a tombstone.

And it’s not an isolated case.

Down to 20–30. Weekly caps hovering around 100–200. Multi-account setups getting flagged within days.

LinkedIn’s detection systems have gotten sharper, faster, and less forgiving.

The message is brutal: if it looks automated, it gets treated like spam.

Doesn’t matter if your content is good. Doesn’t matter if your product is real. If the delivery mechanism is fake, the platform treats the whole package as fake.

They lied.

Posting daily was the golden rule for years. Every LinkedIn guru, every growth hack thread, every “I grew to 50K followers in 90 days” carousel said the same thing: volume wins.

The algorithm doesn’t agree anymore.

LinkedIn’s own guidelines now favour consistency over frequency.

Three posts a week with substance beats five posts a week with filler. The AI models can tell the difference between a post that adds something and a post that exists because you had a calendar slot to fill.

Think about your own feed for a second.

How many posts do you scroll past because they feel like someone ticking a box? “Hiring taught me that people are everything.” “Here are 7 things I learned at a conference.” “Grateful for this journey.”

That’s filler. And the algorithm is getting better at spotting it.

LinkedIn’s AI doesn’t just evaluate content quality. It evaluates content-to-creator alignment.

Let me translate that from platform-speak.

If your posts sound nothing like you if they read like they were written by ChatGPT on a Tuesday afternoon with zero context the algorithm notices. Not because it’s checking for AI specifically. But because your audience engagement patterns shift when the content doesn’t match the human behind it.

Real voice gets real engagement.

Manufactured voice gets pity likes from your mum and Dave from procurement.

Here’s the bit most people miss. LinkedIn hasn’t banned AI content. It doesn’t need to.

It can see the downstream effects of content that doesn’t connect. Low dwell time. Surface-level comments. No shares to DMs. No profile visits after reading.

The algorithm doesn’t detect AI. It detects the absence of resonance.

And resonance requires a real human voice behind the words.

The platform is building a trust score for every creator. Based on patterns: does your content consistently sound like a real person? Do your audience’s comments have depth? Does your engagement look organic or orchestrated?

You can’t hack trust. You can only build it.

Three things. None of them are shortcuts.

Voice. Not “tone of voice” from a brand guidelines PDF that nobody reads. Your actual voice. The way you explain things when you’re not performing. The opinions you hold that make you slightly nervous to share. The stories only you can tell because you lived them.

LinkedIn’s algorithm rewards content that generates genuine conversation. Not “Great post!” comments actual back-and-forth. The kind that happens when someone reads something real and feels compelled to respond with something real.

That requires voice. Not vibes. Not aesthetics. Voice.

Depth over frequency. Two posts a week that make your audience think will outperform five that make them scroll.

The algorithm now measures dwell time how long people spend reading your post before engaging. A 200-word hot take that someone reads in 8 seconds and likes out of habit? Worth less than an 800-word story that keeps someone on the page for two minutes.

Write less. Say more. The maths has changed.

I have clients who post twice a week. That’s it. Two posts. But those posts are built from 90-minute interviews where we pulled out the stuff they actually think not what they think LinkedIn wants to hear. The engagement on two voice-led posts consistently outperforms what they were getting from five generic ones. Every. Single. Time.

Organic engagement patterns. The algorithm is looking for natural audience behaviour. Comments that arrive over hours, not seconds. Engagement from people in your actual industry. Shares to DMs (which LinkedIn can now track) from people who genuinely found value.

Everything fake gets penalised. Everything real gets amplified.

Every algorithm update LinkedIn has made in the last 18 months points in one direction: human content wins.

Not human-ish content. Not AI-generated-then-lightly-edited content. Not “I used a tool to batch my posts for the month” content.

Human content.

The kind that requires someone to sit with a founder for 90 minutes. Extract the way they think. Understand their positioning. Turn their messy expertise into content that sounds like them on their best day.

That’s not a hack. That’s a craft.

And it’s exactly what the algorithm now rewards.

For years, the growth hackers said ghostwriters were unnecessary.

“Just use AI.”

“Just batch your content.”

“Just follow this template.”

“Just join this pod.”

Pods are dead. Automation tools are getting accounts banned. Templates produce filler that the algorithm deprioritises. And AI-generated content the stuff that was supposed to replace human writers triggers what researchers literally call “moral disgust” in audiences.

Meanwhile, the founders who invested in their voice who hired someone to extract it, craft it, and deploy it consistently are the ones the algorithm is now actively promoting.

They didn’t optimise for the algorithm. They optimised for being interesting. And the algorithm finally caught up.

The ghostwriting market is growing 6 - 8% annually. It’s projected to nearly double by 2030. Not because ghostwriters are trendy because the smartest founders worked out years ago that their time is better spent running their company while someone else turns their expertise into content that actually lands.

The ones still debating whether ghostwriting is “authentic” are the same ones still joining engagement pods and wondering why their reach tanked.

“OK but I don’t have time to write 3 high-quality posts a week.”

You don’t have to.

That’s literally what ghostwriters exist for. Not to write FOR you. To write AS you. To take the things you’d say in a meeting or a keynote or a pub rant and turn them into content that sounds like you said it because conceptually, you did.

The algorithm changed. It now penalises everything fake and rewards everything real. Which means the premium isn’t on tools, hacks, or volume.

It’s on voice.

Your voice. Extracted properly. Deployed consistently. Sounding like a human being with something worth saying.

That’s it. That’s the whole strategy.

The algorithm didn’t get smarter. It got honest. The question is whether your content can keep up.

What’s one thing you’d say differently on LinkedIn if you stopped performing and started talking? That’s not rhetorical. I actually want to know.

Written by Sarra Richmond, The Ghost.

I write the posts your favourite founders get credit for. Find me → linkedin.com/in/meetsarra